Resilience-Oriented Distributed Control for Islanded Microgrids: A Lightweight UKF-Enhanced Artificial Rabbit Algorithm

Abstract

This work proposes a resilience-oriented decentralized control scheme for islanded microgrids, coupling the lightweight Unscented Kalman Filter (UKF) with an advanced Artificial Rabbit Algorithm (ARA) to achieve a resilient state estimation system during dynamic operation conditions. The hybrid structure presented achieves the requisite for a computationally efficient means of dealing with distributed energy resources (DERs) whilst guaranteeing an adequate precision for estimation during uncertain electrical disturbance situations. ARA uses the concept of a foraging algorithm, which utilizes biological considerations of animal hunting behavior, to recast estimation in respect to an adaptable noise tuning and sparseness of sigma-point sampling, providing low latency for data reconciliation and real-time motion changing ability. The UKF utilizes a linearized adaptive noise covariance modelling system, which regulates and modifies the confidence of the estimation, altering with respect to changing microgrid system conditions. Furthermore, by using selective sigma-point activation techniques, the overall cost of computation is reduced concurrently with the accurate estimations. The internal relationship between the UKF and the ARA produces an effective closed-loop estimation-control arrangement, thus greatly improving the effect of control on overall stability and control performance. The use of the algorithms presented on the NVIDIA Jetson Orin NX computing hardware system shows an overall benefit in cut transient recovery times of 23% and a reduction in computational strain of 40% when compared to existing centralized control algorithms. The overall use of distributed estimation techniques and adaptable noise tuning means that signal loss by sensors and other communication lapses is more easily withstood. In summary, the control design presented, the UKF-ARA hybrid control situation gives significant gains of operational reliability, computational efficiency and trouble-free operation within the field of microgrid operation. Also being a significant basis for therefore efficient distributed energy management illustrative of greater ease for future developments in resilience distributed energy!.